Are we ready for the automation revolution? New evidence suggests not...yet.

Smart meters regulating household water usage in the US worked a little too well, according to new research.

In the battle against water scarcity, many US cities have implemented regulations limiting how much water residents can use outside in their gardens and yards. But enforcing these restrictions has been challenging, with cities relying on "water cops" to randomly check that homes are following the rules. But what if technology could help us?

Oliver Browne, Ludovica Gazze, Michael Greenstone, and Olga Rostapshova conducted a field experiment in Fresno, California, to explore this possibility.

Working with the city council, their experiment monitored the water use and compliance of about 90,000 households during Fresno's peak outdoor watering season in 2018. The researchers tested whether automated or in-person enforcement methods were more effective by randomising how compliance was monitored. Some households received automated fines when their smart meters recorded excessive water use during restricted hours, others were visited by water cops to check they were following the rules.

The study found that automation significantly improved both enforcement of and compliance with water regulations. In the past, water cops failed to effectively enforce regulations. In the summer of 2016, for instance, 68% of households appear to have violated restrictions, but only 0.4% were sanctioned. During the experiment in 2018, the automated system detected 100% of violations and increased fines to households by 14,200%. If it had been implemented citywide, automation would potentially have fined around 16,000 households each month, resulting in over $2.55 million in fines.

During the experiment, automation also reduced summer water consumption by approximately 3%. If the programme had been implemented citywide, it could have achieved 20% of the annual reductions in residential water use that Governor Gavin Newsom requested in July 2021 to tackle the state's recurring drought. Assuming consistent water conservation efforts, the program could save an estimated 394 million gallons of water annually if rolled out across the city.

These findings highlight the potential of automated enforcement to promote sustainable water usage and help communities meet environmental conservation goals. But the success of automation in catching and punishing rule breakers had a downside. Complaint calls to the city's Department of Public Utilities (DPU) increased drastically: the number of households calling rose by over 650% and complaints and disputes about enforcement actions increased by 1,100%. Responding to the public backlash, the city council relaxed the conservation rules, suspended fines, and eventually outlawed automated enforcement. Some council members drew parallels to ‘Big Brother’ levels of control when discussing the automated enforcement pilot.

The Fresno initiative emphasises the need for policymakers to consider the impact of public perception when introducing automated technologies. Unless policymakers allow for public acceptance and education, applying a perfect system to an imperfect world could be doomed to fail.

References

Gazze, L., Browne, O., Greenstone, M., Rostapshova O. (2022). Man vs. Machine: Technological Promise and Political Limits of Automated Regulation Enforcement. CAGE Working Paper (No. 646).